Michael Barbella, Managing Editor09.06.23
It looked like something Alex Murphy (a.k.a., Robocop) would have worn, but it also closely resembled a wardrobe staple of Tony Stark, Dutch, or The Terminator.
Yet none of those men were the helmet’s rightful owners.
Only Los Angeles-based Daqri could claim proprietorship of the futuristic device, a smart hardhat that provided its owners data about their surroundings through augmented reality technology. Designed for industrial workers, the helmet used cameras and sensors to overlay instructions on complex building and construction equipment.
Despite its groundbreaking potential, Daqri’s $15,000 smart helmet never resonated with customers, nor did its $4,995 smart glasses (designed for professionals). Both inventions were “retired” when the company shut down in September 2019.
Daqri’s demise, however, has not hindered the progression of augmented reality (AR) in various industries, particularly healthcare. AR technology increasingly is being employed by companies to craft a no-fault forward approach to manual assembly and improve production process discipline. Medtech manufacturers are using augmented reality to accelerate collaboration, from conceptualization and development through to assembly line introduction and quality checks. Performing such tasks with the help of AR has been shown to accelerate manufacturing, reduce costs, enhance quality, and improve product safety.
The latter benefit is viable through augmented reality verification. Wixom, Mich.-based LightGuide Inc., for example, offers AR software that uses 3D sensors to ensure each step of the manufacturing process is performed correctly (i.e., proper tool choice, component selection/placement). Similarly, companies can guarantee products are assembled properly by employing high-precision machine vision cameras and other AR sensor technology.
Last year, LightGuide released its smartAR Workstation, a preconfigured, turnkey workstation solution powered by the company’s patented AR digital work instruction software. The smartAR Workstation harnesses digital projection technology, displaying step-by-step work instructions onto a mobile, ergonomic production area. It offers a new way for LightGuide users to integrate AR technology into their factory and deliver immediate impacts on quality and productivity, streamlining and error-proofing manual manufacturing processes.
LightGuide claims its enterprise-ready AR software has helped customers achieve a 90% improvement in quality, 50% improvement in throughput, and a 30% improvement in training efficacy.
“We’ve...witnessed manufacturing technology evolving at an incredible pace,” LightGuide Founder/CEO Paul Ryznar said last November after the firm’s smartAR Workstation won The Assembly Show’s Factory of the Future New Product of the Year award. “This new product drives defect prevention into manual processes and enables manufacturers to rapidly integrate the power of LightGuide AR digital work instructions into their factory environments, unlocking the full potential of their manufacturing workforce.”
Augmented reality is just one way medtech firms are modernizing and streamlining their manufacturing processes. Digital twins, cobots, and artificial intelligence also are transforming traditional assembly and automation tasks—digital twins allow device makers to conduct factory, plant and line simulation and process/workstation simulations, while self-guided cobots bring raw materials from the warehouse to production, move parts between work cells, and deliver finished products.
To better understand the ways in which AR, digital twins, and other technologies are changing medtech assembly and automation processes, Medical Product Outsourcing spoke to more than a half dozen experts over the last several weeks. They included:
Sean Blank: The continuing trend of smaller-sized devices and components has us constantly exploring and overcoming the challenges of handling these smaller pieces. We continue to keep our machines custom to meet the demands.
Crew Feighery: We see continuous improvement in sensing, vision, and robotics technology, coupled with greatly enhanced machine software functionality. Appliances and peripherals for vision systems have improved in quality, dramatically reduced in price, and become more commoditized, which has allowed for utilization where it wouldn’t have been justified in the past. Further the more user-friendly software packages make programming and optimization of these systems much more user-friendly. In an industry where we are constantly working with polymers and qualitative (and sometimes subjective) inspection criteria are needed, these newer software packages allow for progressive improvements in learning over time, reducing the amount of user input required to “train” the systems.
Richard Warren: A core focus from our customers today is the integration of vision technologies in to automated operations. As Gen AI proliferates, the ability to incorporate machine learning in defect detection will allow device manufacturers to make step changes in what was previously subjective visual inspection. Further, the ability to link data sources from production equipment in the analysis and modeling equipment will support advance analytics and improvements to the production system.
Charles Klann: We’re seeing the combination of many different materials into one component to reduce the amount of assembly, as demonstrated by over-molding and two-shot molding.
Craig Occhiato: How much can we analyze in one drop of blood? One area we see new innovations in fluid handling is in automation for micro-dosing. Demands to dose even smaller volumes, quickly, non-contact, with improved accuracy ranging from droplet verification to measuring the smallest volumes. Documenting, storing, and analyzing the data also brings challenges.
Brian Romano: Automation and assembly equipment requires specific skills and maintenance, often which end-user manufacturers don’t have. The leveraging of Industry 4.0 principles to provide remote support quickly and efficiently is what is being required. AGR, for instance, has leveraged seven of the nine pillars of Industry 4.0 to provide data acquisition of both KPIV and machine health variables and the provide remote access support using augmented reality glasses and meeting software. In addition, routine machine and process monitoring that uses AI-enabled software is an optional level of the support program.
Don Stefforia and Tim Witherow: Modular automation. We are utilizing a single base structure and then swapping out the part specific tooling to keep costs down. Multiple fixtures can share a single base, which saves money on future projects. Also, in the retail sector, we have taken note of the larger companies using automation for material/stock handling, which is something that we are keeping our eye on with its potential use in medical plastics when it makes sense.
James Walker: Vision and camera inspection systems are continuing to become more capable, smaller, and less costly, allowing for more versions of in-process inspections, reducing the risk of defects making it to the customer. Characteristics such as color, dimensional accuracy, position, and more can often be checked in milliseconds. This means more characteristics can be verified 100% during high-speed assembly processes. Defects can be detected during the assembly process and rejected before any more work or value can be added to a defective part, saving cost.
Barbella: Customized/personalized solutions have become more popular in recent years. How has this trend affected medtech assembly and/or automation services?
Blank: We have found that it has created more variances of O-rings and seals to be used in these devices. These O-ring and seal variances have not been a challenge for us, however, it has required customers to invest in more tooling for the machines to accommodate all the different styles.
Feighery and Warren: Generally these type of products require very flexible equipment and software design to enable processing of a wide variety of product configuration, while ensuring product quality and output requirements are achieved.
Feighery: If you look back at the origin of each minimally invasive medical device that exists, you will find similarities in the manufacturing process steps, but will rarely see things done identically. The means that there's no one solution that will work for all, requiring customized/personalized solutions for each device. This is the driving factor limiting the larger-scale ability to automate processes, reduce costs, and increase throughput. It cannot be overstated.
Klann: In-house collaboration is a part of all new component design and we are always looking at ways to suggest combining materials within the molding processes to help reduce the need for additional assembly.
Occhiato: As diagnostics and treatments become more patient personalized, I envision more breakthroughs in the ability to tune the manufacturing, assembling, and/or testing of devices using AI and digitization to meet those demands.
Romano: I have no experience or expertise in this area. I don’t believe AGR has needed to provide something that pertains to personalized solutions. However, our machines are routinely set up with change parts, settings, recipes, etc. to change the machine operation to run a product variant. To that degree, AGR machines are customized to run customer specific parts and variants to these.
Stefforia and Witherow: Again, going back to direct part marking, not just for traceability, but also for personalization—automation becomes increasingly more important as volumes get higher.
Walker: For pharmaceutical and drug delivery devices, traceability is a requirement for the customer and the end-user. To ensure traceability, automated processes can be used to serialize production and maintain a database through the lifecycle of a product. On the other hand, automated solutions usually require high volumes of identical product to be a feasible approach. This means if a product requires a customized solution between components or assemblies, it can be challenging to automate. These types of projects can be more suited to manual operations, since operators are much more flexible than automated equipment in general.
Barbella: What market forces are fueling the need for medical device automation?
Blank: The shorter lead time demand to have devices in the market has influenced our workforce to increase both the engineering and machine assembly departments to keep up with the customers’ needs.
Feighery and Warren: The need for manufacturers to provide medical devices of ever-increasing complexity and precision at competitive pricing, coupled with the need for speed to market for new products. Additionally, the labor shortage and skills gap presents a growing challenge for today’s manufacturers. Automation mitigates the labor shortage risk.
Feighery: People, people, people. To elaborate, it is the lack of low-wage operators, which have been taken by warehouse fulfillment jobs (i.e., Amazon) where they can make more money with simpler, less tedious tasks that do not require gowning up and working in a cleanroom all day.
Occhiato: One area we see some opposing forces is in sustainability vs. cost of healthcare. Sustainability is putting new pressure on the throughput/cost calculations. A good example is using disposables (i.e., pipette tips) vs. time and ability to clean. New and improved solutions to advance techniques in non-contact dosing and sensing are helping to reduce the need for disposable tips and time to clean fixed pipette tips.
Romano: One force is the technology changes that drive a shorter product life cycle. This requires a different production and automation outlook.
Stefforia and Witherow: There are two forces fueling the need for medical device automation: labor and demographics. Unfortunately, manual labor is just not as accurate as automation. And as medical devices become more complicated and require increased reliability, automation is increasingly necessary to meet the needs of our customers. As COVID-19 taught us, there’s a strong need to be able to supplement our human capital with automation technology. Regarding demographics—as our population ages globally, there is more need than ever for medical care. A greater demand for medical care will drive a greater demand for medical devices. Once device volumes exceed the threshold of what a person can manage in a timely manner, we must decide to either commit additional labor resources or invest in automation to meet that rising demand.
Walker: Single-use products like medication auto injectors and tamper-evident devices are becoming more common. Since the end-user can only use these types of products one time, the required volumes are much higher than reusable products. Higher volume production often requires automation to meet customer requirements.
Barbella: How are smaller, more complex devices and components challenging medtech assembly and automation technology and services?
Blank: For us, it poses handling challenges such as feeding the smaller parts along with creating and machining tooling required to build the O-ring installation components for our machines. It also has created challenges for our customers to handle the smaller parts themselves.
Feighery and Warren: Smaller parts require innovative approaches using advanced technologies, working with experienced equipment partners in order to develop robust equipment.
Warren: Automation is facilitating the miniaturization and increasing complexity of devices. Smaller parts require innovative approaches using advanced technologies, working with experienced equipment partners to develop robust equipment. Miniaturization of medical devices and components requires the ability to handle and process complex materials with tight tolerances. This precision processing is often integrated with the automation of secondary processing to ensure increased throughput, quality, and compliance.
Klann: Within assembly, more intricate minimally invasive devices are driving the need to combine multiple parts into one component. On the automation side, traditional methods such as bowl feeding components are becoming more challenging due to the reduction in size, leading to micro-robotic feeding and assembly driven by vision system.
Occhiato: Demands to process even smaller volumes of fluid quickly down to single cells, accurately, gently, and within budget will always bring about new challenges demanding more complex devices, smarter components, and real-time diagnostics of the process and/or the patient.
New breakthroughs in solenoid valve coil technology provide diagnostic feedback plus more process monitoring by applying algorithms. Two other areas where we are working on more complex solutions are in microdosing and microfluidics (chip- and cartridge-based technologies). Advances in microplate dosing adding speed, accuracy, direction, and sensing of non-contact droplets. Advances in microfluidics with demands for microsensors to detect bubbles, measure specific ions, pH, conductivity, etc. combined with the latest manufacturing techniques.
Romano: Assembly of smaller devices brings a new set of challenges of having to provide material handling and very accurate and precise assembly machinery and techniques.
Mark Paggioli: There are two things I see as it relates to this. First, technology has been increasing the speed at which the changes to size are occurring and reducing the time in between each iteration of the product, contributing to the shorter product lifecycle. Second, we see more specialization on the supplier side and suppliers with the expertise to produce and provide the subassemblies that are being used.
Stefforia and Witherow: This increasing complexity is driving the need for greater levels of automation. Therefore, it’s incumbent on us to find the best possible process to make a part to the customer’s standard. Once the process is defined, it must be repeated, sometimes thousands or hundreds of thousands of times. These repetitive processes tend to allow humans to become complacent, and that is when mistakes happen. If the process is repeatable and the volumes dictate, generally speaking, any process can be automated. As variation is the enemy of manufacturing, automated assembly allows us to mitigate much of that variation or continue to build in a greater level of robust process control without committing additional labor toward a specific project.
Walker: Smaller device components can be more difficult to fixture and orient in automated equipment. Smaller components are also more easily damaged by automation. More complex assemblies provide more opportunities for something to go wrong during the assembly process, and since there are more operations, there are more opportunities for machine failures.
Barbella: Is there any kind of assembly that cannot be automated? Why?
Blank: The improvements of 3D and vision systems for parts picking has increased the amount of components that can be assembled in an automated system. While that has helped overall, a limiting factor for O-ring installation onto medical devices is if components are too small in size or complex in design to be fed through an automatic system.
Feighery and Warren: Generally, almost everything is possible but if the product volumes are low or if the product design requires a lot of complexity to automate, then automation may not be financially feasible. Design for manufacturing should always be considered early in product design (but almost never is for medical devices). We regularly encounter products that could be automated at far lower cost if manufacturing had been taken into account early on, and often the cost of changing the product design to facilitate manufacturing becomes prohibitive if left too late in the product design cycle.
Feighery: The nature of new product development in medical device is most commonly one of the [following] two scenarios: 1) The R&D team at a larger OEM will work to develop a new device in conjunction with doctors or industry experts with the focus being on functional success of the finished device. How the device is manufactured is typically just several iterations of hand-building a device with in-house R&D lab equipment. In this development, the focus is rarely “what's the best, most repeatable, most efficient way to manufacture this device.” It is far more common that the R&D engineers use whatever tools and process they have on hand or are familiar with. The problem with this is, once the device is finalized, the manufacturing methodologies are locked and will be fixed for essentially the life of the device because the cost, time, and paperwork to change is so significant. 2) Small startup companies develop a novel device for a procedure. With the nature of the startups and the budgets associated, it is common for them to use the least expensive methodologies for manufacture to save costs on CAPEX early on. However, similar to OEMs, once the process is locked it is nearly never able to be changed. These two scenarios represent 90% of the new product development we have seen and are nearly wholely responsible for the challenges faced in automating many of these processes. Design for manufacturing (DFM) is rarely considered until it is too late.
Klann: A risk assessment should be done on any new component to determine the feasibility of automation.
Occhiato: Good question. I think there are steps in assembly and test, especially at startup where human interaction is still necessary, be it a visual inspection or process audits to analyze and possibly adjust parameters.
Romano: The question is probably better asked, “is automation of certain assemblies justifiable with an ROI or ROA?” Almost anything can be automated but the cost of the custom automation for the needed assembly may be prohibitive to implement. Lower-cost robotics and cobots will help bring the number of unjustified assembly projects down but there is still a cost/benefit balance that must be evaluated.
Stefforia and Witherow: In short, no. But it becomes a question about volume and risk assessment. Is there enough volume to merit the capital investment? For example, there are plenty of complex applications that are <500 units per year. In these instances, it wouldn’t make sense to invest in automation because there has to be return on investment. Additionally, due to high risk factors, there are applications where a skilled technician is better suited to evaluate and assemble a part.
Walker: There are two cases in which an assembly is not well suited for automation. One is volume—if the demand is low enough that a manual assembly operation can easily hit target volumes, it can be hard to justify the expense of automation equipment to the customer. The second case is complexity of the assembly in relation to floor space. Manual assembly operations are more flexible, and equipment can be put away to make room for other assembly operations in the same space. Automated assembly equipment usually locks down valuable floor space for one operation, and if that specific operation is not in production, the floor space is not used.
Barbella: What do you think will be possible in medtech assembly and automation technology in the future (say five to 10 years)?
Blank: With the increasing developments just in vision, 3D picking, and 3D parts printing alone, gives the indications that technology will make almost anything possible.
Warren: Adoption of Gen AI will drive increasingly rapid advances in quality control for device manufacturing. Further, these tools can be expanded to more rapidly develop automation solutions and increase the speed with which customized projects can be launched. Increased ability to capture real time manufacturing data coupled with processing of smaller components to tighter tolerances and higher run speeds.
Feighery: The hope will be that new devices are designed for manufacture from the start. However, realistically, with the rate at which things progress in the medical device space due to the FDA and regulations associated, it is unlikely to occur in this time period. We are still stuck with manufacturing processes that were developed 30+ years ago and cannot be changed.
Klann: In seeing how vision systems have impacted assembly and automation technology over the past five years, I would expect the future to look vastly different from today. We would expect increases in the flexibility of production equipment to be able to run multiple different parts due to growth in sensor technology.
Occhiato: I think and hope we will continue to see exciting medical breakthroughs improving our quality of life by adopting more POC and mobile devices moving further to personalized, in-home, and wearable devices. We will see breakthroughs in genomics, with the industry hopefully sharing technologies and findings to speed us to cures and prevention. Medtech assembly and automation will meet the challenge by developing smarter and smaller components and solutions handling much smaller specimens more quickly, sharing and analyzing larger amounts of data.
Romano: Watching the technology change and be adapted in the general population and in assembly and manufacturing, the question becomes twofold. Considering Everett Rogers Diffusion of Innovation theory and Geoffrey Moore’s explanation of the “chasm” gap defining the point at which a product or technology moves into the mainstream, how fast AI technologies will be adopted and accepted into manufacturing, and what regulations may slow or hamper acceptance is one of the points dictating what gets brought forward into assembly technology. The second question concerns which new technologies will be introduced. The availability of robotics on the factory floor at large scale is more predominant today. Do AI and robotics have a joint outcome that provides custom assembly cells that can adapt to an ever-changing environment?
Stefforia and Witherow: Generally speaking, we see a large increase in part traceability requirements coming, like directly marking barcodes onto parts. This has been happening in automotive for about the last 10 years. Also, as computing power increases, there will likely be a need to analyze data. For instance, CT scanning parts for validation will be huge. A 3D model overlaid onto a CAD model is a huge asset when validating. It can only better the quality of the product we are making by taking the guesswork out of CMM vs. check gage/gage pins/micrometers/calipers, etc.
Walker: Additive manufacturing will continue to be used more often in the medical device industry as more materials are developed for human use applications and technology continues to improve. The opportunity to 3D print complete medical devices will reduce the need for secondary automated assembly processes and equipment by allowing for full, articulating assemblies to be printed as one piece. It will also reduce the need for specialized tooling, allowing multiple products to be printed on the same equipment by simply changing the program and/or material used. This can not only eliminate the need for tooling changes but eliminate tooling build times, development times, and tooling storage space requirements.
Yet none of those men were the helmet’s rightful owners.
Only Los Angeles-based Daqri could claim proprietorship of the futuristic device, a smart hardhat that provided its owners data about their surroundings through augmented reality technology. Designed for industrial workers, the helmet used cameras and sensors to overlay instructions on complex building and construction equipment.
Despite its groundbreaking potential, Daqri’s $15,000 smart helmet never resonated with customers, nor did its $4,995 smart glasses (designed for professionals). Both inventions were “retired” when the company shut down in September 2019.
Daqri’s demise, however, has not hindered the progression of augmented reality (AR) in various industries, particularly healthcare. AR technology increasingly is being employed by companies to craft a no-fault forward approach to manual assembly and improve production process discipline. Medtech manufacturers are using augmented reality to accelerate collaboration, from conceptualization and development through to assembly line introduction and quality checks. Performing such tasks with the help of AR has been shown to accelerate manufacturing, reduce costs, enhance quality, and improve product safety.
The latter benefit is viable through augmented reality verification. Wixom, Mich.-based LightGuide Inc., for example, offers AR software that uses 3D sensors to ensure each step of the manufacturing process is performed correctly (i.e., proper tool choice, component selection/placement). Similarly, companies can guarantee products are assembled properly by employing high-precision machine vision cameras and other AR sensor technology.
Last year, LightGuide released its smartAR Workstation, a preconfigured, turnkey workstation solution powered by the company’s patented AR digital work instruction software. The smartAR Workstation harnesses digital projection technology, displaying step-by-step work instructions onto a mobile, ergonomic production area. It offers a new way for LightGuide users to integrate AR technology into their factory and deliver immediate impacts on quality and productivity, streamlining and error-proofing manual manufacturing processes.
LightGuide claims its enterprise-ready AR software has helped customers achieve a 90% improvement in quality, 50% improvement in throughput, and a 30% improvement in training efficacy.
“We’ve...witnessed manufacturing technology evolving at an incredible pace,” LightGuide Founder/CEO Paul Ryznar said last November after the firm’s smartAR Workstation won The Assembly Show’s Factory of the Future New Product of the Year award. “This new product drives defect prevention into manual processes and enables manufacturers to rapidly integrate the power of LightGuide AR digital work instructions into their factory environments, unlocking the full potential of their manufacturing workforce.”
Augmented reality is just one way medtech firms are modernizing and streamlining their manufacturing processes. Digital twins, cobots, and artificial intelligence also are transforming traditional assembly and automation tasks—digital twins allow device makers to conduct factory, plant and line simulation and process/workstation simulations, while self-guided cobots bring raw materials from the warehouse to production, move parts between work cells, and deliver finished products.
To better understand the ways in which AR, digital twins, and other technologies are changing medtech assembly and automation processes, Medical Product Outsourcing spoke to more than a half dozen experts over the last several weeks. They included:
- Sean Blank, sales manager at Automated Industrial Systems Inc., an Erie, Pa.-based custom assembly machine manufacturer
- Crew Feighery, sales vice president - Catheter Technologies, and Richard Warren, chief commercial officer at Medical Manufacturing Technologies, a Charlotte, N.C.-based firm offering process development, applications and equipment, technical solutions, and aftermarket support for medical device companies
- Charles Klann, engineering director at Saint-Gobain, a Portage, Wis.-based silicone-molding manufacturer of medical components
- Craig Occhiato, field segment manager, Burkert Lab Automation and Medical Device Group at Burkert, a global manufacturer of measurement and control systems for liquids and gases
- Brian Romano, technology development director, and Mark Paggioli, marketing and customer service director, at Arthur G. Russell Co. Inc., a Bristol, Conn.-headquartered provider of high-volume medical device production assembly systems and automated assembly machines
- Don Stefforia, manufacturing engineering manager, and Tim Witherow, assistant plant manager, at PTI Engineered Plastics Inc., an injection molder and manufacturer of plastic components and assemblies based in Macomb, Mich.
- James Walker, process development engineer – Assembly/Design, at Raumedic Inc., a Mills River, N.C.-headquartered firm that designs, develops, and produces polymer- and silicone-based solutions for customer-specific medical and pharmaceutical applications
Sean Blank: The continuing trend of smaller-sized devices and components has us constantly exploring and overcoming the challenges of handling these smaller pieces. We continue to keep our machines custom to meet the demands.
Crew Feighery: We see continuous improvement in sensing, vision, and robotics technology, coupled with greatly enhanced machine software functionality. Appliances and peripherals for vision systems have improved in quality, dramatically reduced in price, and become more commoditized, which has allowed for utilization where it wouldn’t have been justified in the past. Further the more user-friendly software packages make programming and optimization of these systems much more user-friendly. In an industry where we are constantly working with polymers and qualitative (and sometimes subjective) inspection criteria are needed, these newer software packages allow for progressive improvements in learning over time, reducing the amount of user input required to “train” the systems.
Richard Warren: A core focus from our customers today is the integration of vision technologies in to automated operations. As Gen AI proliferates, the ability to incorporate machine learning in defect detection will allow device manufacturers to make step changes in what was previously subjective visual inspection. Further, the ability to link data sources from production equipment in the analysis and modeling equipment will support advance analytics and improvements to the production system.
Charles Klann: We’re seeing the combination of many different materials into one component to reduce the amount of assembly, as demonstrated by over-molding and two-shot molding.
Craig Occhiato: How much can we analyze in one drop of blood? One area we see new innovations in fluid handling is in automation for micro-dosing. Demands to dose even smaller volumes, quickly, non-contact, with improved accuracy ranging from droplet verification to measuring the smallest volumes. Documenting, storing, and analyzing the data also brings challenges.
Brian Romano: Automation and assembly equipment requires specific skills and maintenance, often which end-user manufacturers don’t have. The leveraging of Industry 4.0 principles to provide remote support quickly and efficiently is what is being required. AGR, for instance, has leveraged seven of the nine pillars of Industry 4.0 to provide data acquisition of both KPIV and machine health variables and the provide remote access support using augmented reality glasses and meeting software. In addition, routine machine and process monitoring that uses AI-enabled software is an optional level of the support program.
Don Stefforia and Tim Witherow: Modular automation. We are utilizing a single base structure and then swapping out the part specific tooling to keep costs down. Multiple fixtures can share a single base, which saves money on future projects. Also, in the retail sector, we have taken note of the larger companies using automation for material/stock handling, which is something that we are keeping our eye on with its potential use in medical plastics when it makes sense.
James Walker: Vision and camera inspection systems are continuing to become more capable, smaller, and less costly, allowing for more versions of in-process inspections, reducing the risk of defects making it to the customer. Characteristics such as color, dimensional accuracy, position, and more can often be checked in milliseconds. This means more characteristics can be verified 100% during high-speed assembly processes. Defects can be detected during the assembly process and rejected before any more work or value can be added to a defective part, saving cost.
Barbella: Customized/personalized solutions have become more popular in recent years. How has this trend affected medtech assembly and/or automation services?
Blank: We have found that it has created more variances of O-rings and seals to be used in these devices. These O-ring and seal variances have not been a challenge for us, however, it has required customers to invest in more tooling for the machines to accommodate all the different styles.
Feighery and Warren: Generally these type of products require very flexible equipment and software design to enable processing of a wide variety of product configuration, while ensuring product quality and output requirements are achieved.
Feighery: If you look back at the origin of each minimally invasive medical device that exists, you will find similarities in the manufacturing process steps, but will rarely see things done identically. The means that there's no one solution that will work for all, requiring customized/personalized solutions for each device. This is the driving factor limiting the larger-scale ability to automate processes, reduce costs, and increase throughput. It cannot be overstated.
Klann: In-house collaboration is a part of all new component design and we are always looking at ways to suggest combining materials within the molding processes to help reduce the need for additional assembly.
Occhiato: As diagnostics and treatments become more patient personalized, I envision more breakthroughs in the ability to tune the manufacturing, assembling, and/or testing of devices using AI and digitization to meet those demands.
Romano: I have no experience or expertise in this area. I don’t believe AGR has needed to provide something that pertains to personalized solutions. However, our machines are routinely set up with change parts, settings, recipes, etc. to change the machine operation to run a product variant. To that degree, AGR machines are customized to run customer specific parts and variants to these.
Stefforia and Witherow: Again, going back to direct part marking, not just for traceability, but also for personalization—automation becomes increasingly more important as volumes get higher.
Walker: For pharmaceutical and drug delivery devices, traceability is a requirement for the customer and the end-user. To ensure traceability, automated processes can be used to serialize production and maintain a database through the lifecycle of a product. On the other hand, automated solutions usually require high volumes of identical product to be a feasible approach. This means if a product requires a customized solution between components or assemblies, it can be challenging to automate. These types of projects can be more suited to manual operations, since operators are much more flexible than automated equipment in general.
Barbella: What market forces are fueling the need for medical device automation?
Blank: The shorter lead time demand to have devices in the market has influenced our workforce to increase both the engineering and machine assembly departments to keep up with the customers’ needs.
Feighery and Warren: The need for manufacturers to provide medical devices of ever-increasing complexity and precision at competitive pricing, coupled with the need for speed to market for new products. Additionally, the labor shortage and skills gap presents a growing challenge for today’s manufacturers. Automation mitigates the labor shortage risk.
Feighery: People, people, people. To elaborate, it is the lack of low-wage operators, which have been taken by warehouse fulfillment jobs (i.e., Amazon) where they can make more money with simpler, less tedious tasks that do not require gowning up and working in a cleanroom all day.
Occhiato: One area we see some opposing forces is in sustainability vs. cost of healthcare. Sustainability is putting new pressure on the throughput/cost calculations. A good example is using disposables (i.e., pipette tips) vs. time and ability to clean. New and improved solutions to advance techniques in non-contact dosing and sensing are helping to reduce the need for disposable tips and time to clean fixed pipette tips.
Romano: One force is the technology changes that drive a shorter product life cycle. This requires a different production and automation outlook.
Stefforia and Witherow: There are two forces fueling the need for medical device automation: labor and demographics. Unfortunately, manual labor is just not as accurate as automation. And as medical devices become more complicated and require increased reliability, automation is increasingly necessary to meet the needs of our customers. As COVID-19 taught us, there’s a strong need to be able to supplement our human capital with automation technology. Regarding demographics—as our population ages globally, there is more need than ever for medical care. A greater demand for medical care will drive a greater demand for medical devices. Once device volumes exceed the threshold of what a person can manage in a timely manner, we must decide to either commit additional labor resources or invest in automation to meet that rising demand.
Walker: Single-use products like medication auto injectors and tamper-evident devices are becoming more common. Since the end-user can only use these types of products one time, the required volumes are much higher than reusable products. Higher volume production often requires automation to meet customer requirements.
Barbella: How are smaller, more complex devices and components challenging medtech assembly and automation technology and services?
Blank: For us, it poses handling challenges such as feeding the smaller parts along with creating and machining tooling required to build the O-ring installation components for our machines. It also has created challenges for our customers to handle the smaller parts themselves.
Feighery and Warren: Smaller parts require innovative approaches using advanced technologies, working with experienced equipment partners in order to develop robust equipment.
Warren: Automation is facilitating the miniaturization and increasing complexity of devices. Smaller parts require innovative approaches using advanced technologies, working with experienced equipment partners to develop robust equipment. Miniaturization of medical devices and components requires the ability to handle and process complex materials with tight tolerances. This precision processing is often integrated with the automation of secondary processing to ensure increased throughput, quality, and compliance.
Klann: Within assembly, more intricate minimally invasive devices are driving the need to combine multiple parts into one component. On the automation side, traditional methods such as bowl feeding components are becoming more challenging due to the reduction in size, leading to micro-robotic feeding and assembly driven by vision system.
Occhiato: Demands to process even smaller volumes of fluid quickly down to single cells, accurately, gently, and within budget will always bring about new challenges demanding more complex devices, smarter components, and real-time diagnostics of the process and/or the patient.
New breakthroughs in solenoid valve coil technology provide diagnostic feedback plus more process monitoring by applying algorithms. Two other areas where we are working on more complex solutions are in microdosing and microfluidics (chip- and cartridge-based technologies). Advances in microplate dosing adding speed, accuracy, direction, and sensing of non-contact droplets. Advances in microfluidics with demands for microsensors to detect bubbles, measure specific ions, pH, conductivity, etc. combined with the latest manufacturing techniques.
Romano: Assembly of smaller devices brings a new set of challenges of having to provide material handling and very accurate and precise assembly machinery and techniques.
Mark Paggioli: There are two things I see as it relates to this. First, technology has been increasing the speed at which the changes to size are occurring and reducing the time in between each iteration of the product, contributing to the shorter product lifecycle. Second, we see more specialization on the supplier side and suppliers with the expertise to produce and provide the subassemblies that are being used.
Stefforia and Witherow: This increasing complexity is driving the need for greater levels of automation. Therefore, it’s incumbent on us to find the best possible process to make a part to the customer’s standard. Once the process is defined, it must be repeated, sometimes thousands or hundreds of thousands of times. These repetitive processes tend to allow humans to become complacent, and that is when mistakes happen. If the process is repeatable and the volumes dictate, generally speaking, any process can be automated. As variation is the enemy of manufacturing, automated assembly allows us to mitigate much of that variation or continue to build in a greater level of robust process control without committing additional labor toward a specific project.
Walker: Smaller device components can be more difficult to fixture and orient in automated equipment. Smaller components are also more easily damaged by automation. More complex assemblies provide more opportunities for something to go wrong during the assembly process, and since there are more operations, there are more opportunities for machine failures.
Barbella: Is there any kind of assembly that cannot be automated? Why?
Blank: The improvements of 3D and vision systems for parts picking has increased the amount of components that can be assembled in an automated system. While that has helped overall, a limiting factor for O-ring installation onto medical devices is if components are too small in size or complex in design to be fed through an automatic system.
Feighery and Warren: Generally, almost everything is possible but if the product volumes are low or if the product design requires a lot of complexity to automate, then automation may not be financially feasible. Design for manufacturing should always be considered early in product design (but almost never is for medical devices). We regularly encounter products that could be automated at far lower cost if manufacturing had been taken into account early on, and often the cost of changing the product design to facilitate manufacturing becomes prohibitive if left too late in the product design cycle.
Feighery: The nature of new product development in medical device is most commonly one of the [following] two scenarios: 1) The R&D team at a larger OEM will work to develop a new device in conjunction with doctors or industry experts with the focus being on functional success of the finished device. How the device is manufactured is typically just several iterations of hand-building a device with in-house R&D lab equipment. In this development, the focus is rarely “what's the best, most repeatable, most efficient way to manufacture this device.” It is far more common that the R&D engineers use whatever tools and process they have on hand or are familiar with. The problem with this is, once the device is finalized, the manufacturing methodologies are locked and will be fixed for essentially the life of the device because the cost, time, and paperwork to change is so significant. 2) Small startup companies develop a novel device for a procedure. With the nature of the startups and the budgets associated, it is common for them to use the least expensive methodologies for manufacture to save costs on CAPEX early on. However, similar to OEMs, once the process is locked it is nearly never able to be changed. These two scenarios represent 90% of the new product development we have seen and are nearly wholely responsible for the challenges faced in automating many of these processes. Design for manufacturing (DFM) is rarely considered until it is too late.
Klann: A risk assessment should be done on any new component to determine the feasibility of automation.
Occhiato: Good question. I think there are steps in assembly and test, especially at startup where human interaction is still necessary, be it a visual inspection or process audits to analyze and possibly adjust parameters.
Romano: The question is probably better asked, “is automation of certain assemblies justifiable with an ROI or ROA?” Almost anything can be automated but the cost of the custom automation for the needed assembly may be prohibitive to implement. Lower-cost robotics and cobots will help bring the number of unjustified assembly projects down but there is still a cost/benefit balance that must be evaluated.
Stefforia and Witherow: In short, no. But it becomes a question about volume and risk assessment. Is there enough volume to merit the capital investment? For example, there are plenty of complex applications that are <500 units per year. In these instances, it wouldn’t make sense to invest in automation because there has to be return on investment. Additionally, due to high risk factors, there are applications where a skilled technician is better suited to evaluate and assemble a part.
Walker: There are two cases in which an assembly is not well suited for automation. One is volume—if the demand is low enough that a manual assembly operation can easily hit target volumes, it can be hard to justify the expense of automation equipment to the customer. The second case is complexity of the assembly in relation to floor space. Manual assembly operations are more flexible, and equipment can be put away to make room for other assembly operations in the same space. Automated assembly equipment usually locks down valuable floor space for one operation, and if that specific operation is not in production, the floor space is not used.
Barbella: What do you think will be possible in medtech assembly and automation technology in the future (say five to 10 years)?
Blank: With the increasing developments just in vision, 3D picking, and 3D parts printing alone, gives the indications that technology will make almost anything possible.
Warren: Adoption of Gen AI will drive increasingly rapid advances in quality control for device manufacturing. Further, these tools can be expanded to more rapidly develop automation solutions and increase the speed with which customized projects can be launched. Increased ability to capture real time manufacturing data coupled with processing of smaller components to tighter tolerances and higher run speeds.
Feighery: The hope will be that new devices are designed for manufacture from the start. However, realistically, with the rate at which things progress in the medical device space due to the FDA and regulations associated, it is unlikely to occur in this time period. We are still stuck with manufacturing processes that were developed 30+ years ago and cannot be changed.
Klann: In seeing how vision systems have impacted assembly and automation technology over the past five years, I would expect the future to look vastly different from today. We would expect increases in the flexibility of production equipment to be able to run multiple different parts due to growth in sensor technology.
Occhiato: I think and hope we will continue to see exciting medical breakthroughs improving our quality of life by adopting more POC and mobile devices moving further to personalized, in-home, and wearable devices. We will see breakthroughs in genomics, with the industry hopefully sharing technologies and findings to speed us to cures and prevention. Medtech assembly and automation will meet the challenge by developing smarter and smaller components and solutions handling much smaller specimens more quickly, sharing and analyzing larger amounts of data.
Romano: Watching the technology change and be adapted in the general population and in assembly and manufacturing, the question becomes twofold. Considering Everett Rogers Diffusion of Innovation theory and Geoffrey Moore’s explanation of the “chasm” gap defining the point at which a product or technology moves into the mainstream, how fast AI technologies will be adopted and accepted into manufacturing, and what regulations may slow or hamper acceptance is one of the points dictating what gets brought forward into assembly technology. The second question concerns which new technologies will be introduced. The availability of robotics on the factory floor at large scale is more predominant today. Do AI and robotics have a joint outcome that provides custom assembly cells that can adapt to an ever-changing environment?
Stefforia and Witherow: Generally speaking, we see a large increase in part traceability requirements coming, like directly marking barcodes onto parts. This has been happening in automotive for about the last 10 years. Also, as computing power increases, there will likely be a need to analyze data. For instance, CT scanning parts for validation will be huge. A 3D model overlaid onto a CAD model is a huge asset when validating. It can only better the quality of the product we are making by taking the guesswork out of CMM vs. check gage/gage pins/micrometers/calipers, etc.
Walker: Additive manufacturing will continue to be used more often in the medical device industry as more materials are developed for human use applications and technology continues to improve. The opportunity to 3D print complete medical devices will reduce the need for secondary automated assembly processes and equipment by allowing for full, articulating assemblies to be printed as one piece. It will also reduce the need for specialized tooling, allowing multiple products to be printed on the same equipment by simply changing the program and/or material used. This can not only eliminate the need for tooling changes but eliminate tooling build times, development times, and tooling storage space requirements.